Response to consultation on Guidelines on credit risk management practices and accounting for expected credit losses
Go back
The way to achieve this in a thorough and consistent way, is to align wording in the final guidelines to existing practise and definitions from IFRS. This includes reflecting that both proportionality and materiality principles must be considered together.
We agree that the proposed guidelines should be applied in a manner that is proportionate to the size, internal organisation, nature, scope and complexity of a credit institutions activities. Applying proportionality should however not be limited to less complex banks, as smaller subsidiaries, branches or immaterial portfolios of large credit institutions may face similar implementation issues and therefore also benefit from applying the proportionality principles. We therefore suggest that further guidance is provided around how such assessments of proportionality are to be performed and applied. In the absence of further guidance we are concerned that proportionality assessments may not be applied in a consistent manner across credit institutions and/or over time.
The concept of materiality is a fundamental principle of all financial reporting and embedded within the IASB Conceptual Framework. It needs to be clear therefore that the discussion of materiality in the proposed guidelines does not override the concept of materiality in IFRS and that the materiality thresholds applicable under IFRS should also be applied to the proposed guidelines.
It is important that the guidelines avoid taking an unnecessarily prescriptive approach where this is not aligned with IFRS9. There are certain instances where the BCBS guidance provides for a more granular tick-box approach and goes against the IFRB principles. We provide specific examples in our response to Question 6.
Redefining terms such as Reasonable and supportable information" is inappropriate as these terms exist within Accounting Standards. This could create operational complexity and lead to confusion.
Para 21, page 17 Guidance on the consideration of reasonable and supportable information is already discussed within IFRS 9 and should not be repeated within these EBA guidelines.
Again, we caution against adding definitions in these guidelines that are already established and clear. This would avoid potentially introducing conflicting interpretations as well as simplifying the guidelines.
Para 36 a-h,
page 23
These procedures already exist in the IFRS 9 requirements. Therefore we do not see any value in these being repeated in these guidelines.
It is unclear why requirements on content on lending policies are listed under sound ECL methodologies. Not only is this potentially confusing but these requirements are already established and common practise. We reiterate our view that repeating or duplicating existing provisions and requirements could lead to uncertainty and overly complex guidelines.
While the benefits of the proposed guidelines are particularly clear for large credit institutions, for smaller credit institutions, immaterial subsidiaries/branches/portfolios, it would be helpful to have a further clarification around how the implementation benefits outweigh the costs.
Para 38e,
page 25
We do not consider credit default swap spreads as market indicators of future performance. We suggest this reference be removed.
Para 41,
page 26
We suggest the guidelines avoid making specific references to actions of lending staff. These guidelines are not the correct place for statements such as "lending staff to promptly notify the institution’s accounting function when exposures are renegotiated or modified to ensure appropriate accounting for the change. For more complex renegotiations and modifications, regular communication between the lending staff and the accounting function should take place."
These aspects are covered under the overall control and processes framework. These guidelines should maintain the IFRS approach of providing principles but not seek to provide such granular mandating.
Para 64,
page 30
This section covers model development requirements which should not be listed as validation requirements. This is potentially confusing and misleading.
Para 67 b, first bullet,
page 31
As per the comment on para 64, this is model development and should be listed in the guidelines accordingly.
67 b, second bullet,
page 32
The requirement to demonstrate conceptual soundness and appropriateness is not to be fulfilled by the validation function but by the development function.
The validation function is an internal independent function to assess the conceptual soundness but not to demonstrate it.
67 b, third bullet,
page 32
We recommend deleting the requirement for fixed thresholds for assessing model performance. It is not always feasible to determine appropriate thresholds – specifically in instances of small portfolios or in cases of statistically insignificant databases
It should also be acknowledged that remedial actions are specific to the identified issue and recalibrating or overhauling the model is not always necessary.
Para 68,
page 32
We believe that the term "robust" should be replaced by "reasonable and supportable". This should be aligned with paragraph 131 (4.3.3. use of practical expedients) where guidance is given using the term "reasonable and supportable".
Para 78 - 85,
page 34
.
We suggest that disclosure requirements be left to accounting and supervisory frameworks. Where the EBA feels disclosures over and above those already mandated by IFRS and supervisory regimes are necessary, it should be explained why this is deemed the case.
Para 86-89,
page 35
Para 86: The guidelines should avoid suggesting whether Expected Credit Loss (ECL) allowances will be nil or not. Each credit institution will have its own lending exposures suitable for its risk appetite. Additionally, an ECL of nil is hard to explain - particularly if they are determined on a complete statistical approach via Loss Given Default /Probability of Default and not individually assessed. An ECL of nil would mean never having observed any loss on this group of loans and even taking into account every reasonable forward looking outcome there is zero possibility losses will ever occur.
Para 87: We would appreciate further guidance on what "Credit institutions should adopt an active approach to assessing and measuring 12-month ECL" means?
Para 88: We do not believe these guidelines should seek to explain IFRS 9 and the basis for IFRS 9 principles. This could be confusing and may risk divergent interpretations.
Para 94,
page 37
Frequent segmentation and re-segmentation of portfolios could lead to statistical data for the re-segmentation portfolios being difficult to verify appropriately.
Para 107 a-f,
page 41
While these indicators are useful, the guidelines should not prescribe a methodology which may be interpreted to require credit institutions to give greater prominence and weighting to specific indicators when assessing significant increase in credit risk, rather than considering all relevant information. This is not consistent with IFRS 9 B5.5.15 which requires considering reasonable and supportable information, with a non-exhaustive list of information which may be relevant, without giving prominence to any specific factors. Our suggestion is that the proposed guidelines cross-reference IFRS 9 principles where relevant
Additionally we do not share the view that loan pricing is a suitable indicator of an increase in a client's credit risk as loan pricing may be impacted by factors not related to the borrower's credit risk (e.g. a credit institutions cost of funding).
Para 108- 126,
page 41-45
The guidelines are overly prescriptive is providing an extensive list of factors to be considered when determining a significant increase in credit risk for lending exposures. This contradicts the principle based nature of IFRS 9. We suggest the guidelines avoid such an approach."
Question 1: Is the scope of application of the guidelines appropriate and sufficiently clear?
Some further clarification is required. Paragraph 11, page 13 states that these guidelines should be applied on an individual, sub-consolidated and consolidated basis. However, in certain circumstances there may be non-EU subsidiaries of EU banks that would be subject to local expected credit loss rules. Where this happens, we would welcome clarification that non-EU subsidiaries would be expected to follow local rules rather than these guidelines.Question 3: Please provide any comments you may have on the appropriateness of the proposed proportionality approach.
Whilst the guidelines provide information on the definition and treatment of material institutions and portfolios, they do not address the issue of noncomplex or immaterial portfolios. These should explicitly be considered in the context of the proportionality principle.The way to achieve this in a thorough and consistent way, is to align wording in the final guidelines to existing practise and definitions from IFRS. This includes reflecting that both proportionality and materiality principles must be considered together.
We agree that the proposed guidelines should be applied in a manner that is proportionate to the size, internal organisation, nature, scope and complexity of a credit institutions activities. Applying proportionality should however not be limited to less complex banks, as smaller subsidiaries, branches or immaterial portfolios of large credit institutions may face similar implementation issues and therefore also benefit from applying the proportionality principles. We therefore suggest that further guidance is provided around how such assessments of proportionality are to be performed and applied. In the absence of further guidance we are concerned that proportionality assessments may not be applied in a consistent manner across credit institutions and/or over time.
The concept of materiality is a fundamental principle of all financial reporting and embedded within the IASB Conceptual Framework. It needs to be clear therefore that the discussion of materiality in the proposed guidelines does not override the concept of materiality in IFRS and that the materiality thresholds applicable under IFRS should also be applied to the proposed guidelines.
Question 4: Do you agree with the draft guidelines which introduce the relevant BCBS Guidance in the EU regulatory framework? Are there additional issues for which the EBA Guidelines should be amended in the context of finalising the guidelines?
We would appreciate confirmation that EBA Guidelines will supersede BCBS guidance for EU banks.It is important that the guidelines avoid taking an unnecessarily prescriptive approach where this is not aligned with IFRS9. There are certain instances where the BCBS guidance provides for a more granular tick-box approach and goes against the IFRB principles. We provide specific examples in our response to Question 6.
Question 6: Please provide any additional comments on the draft guidelines.
Page 14, definitions The Guidelines should limit definitions to terms that are uniquely used within the Consultation Paper.Redefining terms such as Reasonable and supportable information" is inappropriate as these terms exist within Accounting Standards. This could create operational complexity and lead to confusion.
Para 21, page 17 Guidance on the consideration of reasonable and supportable information is already discussed within IFRS 9 and should not be repeated within these EBA guidelines.
Again, we caution against adding definitions in these guidelines that are already established and clear. This would avoid potentially introducing conflicting interpretations as well as simplifying the guidelines.
Para 36 a-h,
page 23
These procedures already exist in the IFRS 9 requirements. Therefore we do not see any value in these being repeated in these guidelines.
It is unclear why requirements on content on lending policies are listed under sound ECL methodologies. Not only is this potentially confusing but these requirements are already established and common practise. We reiterate our view that repeating or duplicating existing provisions and requirements could lead to uncertainty and overly complex guidelines.
While the benefits of the proposed guidelines are particularly clear for large credit institutions, for smaller credit institutions, immaterial subsidiaries/branches/portfolios, it would be helpful to have a further clarification around how the implementation benefits outweigh the costs.
Para 38e,
page 25
We do not consider credit default swap spreads as market indicators of future performance. We suggest this reference be removed.
Para 41,
page 26
We suggest the guidelines avoid making specific references to actions of lending staff. These guidelines are not the correct place for statements such as "lending staff to promptly notify the institution’s accounting function when exposures are renegotiated or modified to ensure appropriate accounting for the change. For more complex renegotiations and modifications, regular communication between the lending staff and the accounting function should take place."
These aspects are covered under the overall control and processes framework. These guidelines should maintain the IFRS approach of providing principles but not seek to provide such granular mandating.
Para 64,
page 30
This section covers model development requirements which should not be listed as validation requirements. This is potentially confusing and misleading.
Para 67 b, first bullet,
page 31
As per the comment on para 64, this is model development and should be listed in the guidelines accordingly.
67 b, second bullet,
page 32
The requirement to demonstrate conceptual soundness and appropriateness is not to be fulfilled by the validation function but by the development function.
The validation function is an internal independent function to assess the conceptual soundness but not to demonstrate it.
67 b, third bullet,
page 32
We recommend deleting the requirement for fixed thresholds for assessing model performance. It is not always feasible to determine appropriate thresholds – specifically in instances of small portfolios or in cases of statistically insignificant databases
It should also be acknowledged that remedial actions are specific to the identified issue and recalibrating or overhauling the model is not always necessary.
Para 68,
page 32
We believe that the term "robust" should be replaced by "reasonable and supportable". This should be aligned with paragraph 131 (4.3.3. use of practical expedients) where guidance is given using the term "reasonable and supportable".
Para 78 - 85,
page 34
.
We suggest that disclosure requirements be left to accounting and supervisory frameworks. Where the EBA feels disclosures over and above those already mandated by IFRS and supervisory regimes are necessary, it should be explained why this is deemed the case.
Para 86-89,
page 35
Para 86: The guidelines should avoid suggesting whether Expected Credit Loss (ECL) allowances will be nil or not. Each credit institution will have its own lending exposures suitable for its risk appetite. Additionally, an ECL of nil is hard to explain - particularly if they are determined on a complete statistical approach via Loss Given Default /Probability of Default and not individually assessed. An ECL of nil would mean never having observed any loss on this group of loans and even taking into account every reasonable forward looking outcome there is zero possibility losses will ever occur.
Para 87: We would appreciate further guidance on what "Credit institutions should adopt an active approach to assessing and measuring 12-month ECL" means?
Para 88: We do not believe these guidelines should seek to explain IFRS 9 and the basis for IFRS 9 principles. This could be confusing and may risk divergent interpretations.
Para 94,
page 37
Frequent segmentation and re-segmentation of portfolios could lead to statistical data for the re-segmentation portfolios being difficult to verify appropriately.
Para 107 a-f,
page 41
While these indicators are useful, the guidelines should not prescribe a methodology which may be interpreted to require credit institutions to give greater prominence and weighting to specific indicators when assessing significant increase in credit risk, rather than considering all relevant information. This is not consistent with IFRS 9 B5.5.15 which requires considering reasonable and supportable information, with a non-exhaustive list of information which may be relevant, without giving prominence to any specific factors. Our suggestion is that the proposed guidelines cross-reference IFRS 9 principles where relevant
Additionally we do not share the view that loan pricing is a suitable indicator of an increase in a client's credit risk as loan pricing may be impacted by factors not related to the borrower's credit risk (e.g. a credit institutions cost of funding).
Para 108- 126,
page 41-45
The guidelines are overly prescriptive is providing an extensive list of factors to be considered when determining a significant increase in credit risk for lending exposures. This contradicts the principle based nature of IFRS 9. We suggest the guidelines avoid such an approach."